JPH0568169B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a training simulator, and in particular is used to train operators who operate an electric power system.

Conventionally, a power flow calculation system with the hardware configuration shown in FIG. 1 has been used for operational calculations of power systems. In the figure, 1 is a digital computer, 2
is a disk device used as an external storage device,
3 is a system typewriter, 4 is a console used as a man-machine interface (including a CRT display), 5 is a card reader,
Reference numeral 6 indicates a line printer, and 7 indicates a system monitoring panel.

In the conventional power flow calculation system shown in Fig. 1, the system status is set using an input device such as a CRT display on the operation console 4, a program in the digital computer 1 is started, and as soon as the program is finished, the CRT display etc. Display the output results on the output device.

For example, if the power system changes from the state in Figure 2A to Figure 2B
Consider the case where the state changes to .

In Figure 2, the power system consists of two generators 11
A transmission line 13ij connects buses 12i and 12j of buses 12i and 11j, and buses 12i and 12j
j and the bus bar 15 of the load 14 are connected by power transmission lines 16i and 16j, respectively. However, when the switching devices 17i and 17j provided on the power transmission line 13ij between the bus bars 12i and 12j are opened from the state shown in FIG. 2A and the state shown in FIG. calculated the system power flow according to the procedure shown in Figure 3.

There are various power flow calculation methods, but in the case of the Newton-Raphson method using the node equation, the principle of the convergence calculation to determine the bus voltage and phase angle is as follows: Δθ ΔV _i+1 = J ₁₂ J ₁₂ J ₂₁ J _{22-1 i} ΔP ΔQ _i (1) As shown in (1), the correction amounts ΔP and ΔQ are calculated using θ and V of the i-th repetition number. Here, Δθ is the phase angle correction amount, ΔV is the voltage correction amount,
ΔP is the active power correction amount, ΔQ is the reactive power correction amount, j _ij is the Jacobian matrix, and J ₁₁ = ∂P/∂θ, J ₁₂ = P/∂V J ₂₁ = ∂Q/∂θ, J ₂₂ = ∂ Q/∂V …(2) is shown. In this way, based on equation (1), the correction amount ΔP,
Compute ΔQ, then compute the Jacobian matrix J _ij ,
Find its inverse matrix J ^-1 . Next, J ^-1 and [ΔP, ΔQ]
Find [Δθ, ΔV] ^T by taking the product of ^T.
This calculation result is used again as a correction amount in the next calculation, and the above calculation is thus repeated until θ and V converge. If θ and V eventually converge,
Using these values, power flows P _ij and Q _ij of branches (including power transmission lines, transformer equipment, etc.) are calculated. A program within a digital computer processes the principles described above.

Next, the operation of the conventional power system operation training simulator based on the power flow calculation method using the Jacobian matrix described above will be explained with reference to the flowchart shown in FIG.

Step S1: Start up the simulation program in the digital computer 1.

Step S2: Prior to execution of the simulation program, a preprocessing program is executed and the second
The system status of the power system diagram shown in Figure B, the operation status, the on/off status of the disconnectors and disconnectors 17 _i and 17 _j , and the output disconnection load 14 of the generators 11 _i and 11 _j are set in the program. In this way, before executing power flow calculations, a separation check, etc. is performed, and the system configuration to be calculated is recognized.

Step S3: Initial values of voltage V and phase angle θ of each bus bar 12 _i , 12 _j , 15 are calculated.

Step S4: Correction amount, that is, ΔP (active power correction amount) and ΔQ (reactive power correction amount) shown in equation (1).
perform the calculation.

Step S5: Based on each of the above correction amounts ΔP and ΔQ
According to equation (2), the Jacobian matrix J and its inverse matrix J ^-1
seek.

Step S6: Calculate voltage correction amount ΔV and phase angle correction amount Δθ from the inverse matrix J ^-1 and correction amounts ΔP and ΔQ according to equation (1).

Step S7: This calculation result is used again as a correction amount, and steps S4 to S7 are repeated until Δθ and ΔV converge in equation (1).

Step S8: Δθ and ΔV at step S7
If the judgment result that the has converged is obtained, the convergence value θ,
Using V, calculate the power flow of the power system diagram shown in Figure 2B.
P _ij and Q _ij are calculated, and the program is terminated in step S9.

In this way, in the conventional training simulator, as shown in the flowchart shown in FIG. 3, regardless of whether the system configuration has changed or not, the computer 1 calculates the initial value in step S3,
The calculation of the Jacobian inverse matrix J ^-1 at step S5 is
The calculation will be performed every time the program is started. However, this method has the drawback of increasing the calculation execution time and increasing the computer occupation rate.

Also, each time the calculation is repeated, the Jacobian inverse matrix is
J ^-1 had to be calculated, and the cycle for displaying the power flow calculation results on the CTR in the simulator's system monitoring panel became long, creating a problem that it was not possible to simulate the power flow of the power system in real time. .

This invention was made to solve the above problems, and by speeding up power flow calculations even more than conventional power flow calculation systems, it is possible to perform complex power flow calculations that correspond to the behavior of a large number of power devices and loads. The purpose of the present invention is to provide a training simulator that can be executed almost in real time even when calculation processing such as data processing is required.

An example of the present invention will be described in detail below with reference to the drawings.
The hardware configuration of the training simulator is the same as that of the conventional power flow calculation system as described above with reference to FIG.
A power flow calculation program with the calculation procedure shown in FIG. 4 is implemented in the system.

The power flow calculation program of the training simulator in this embodiment will be explained according to the flowchart of FIG.

First, the system status and operation status of the power system diagram that is the target of this power flow calculation, the disconnector, and the disconnector 17 _i , 1
_7. The on/off state of the generator 11 _i , 11 _j and the disconnection load 14 are set in the program as shown in FIG. 2A.

In Figure 2A, the power system consists of two generators 1
Transmission line 13 runs between busbars 12 _i and 12 _j of 1 _i and 11 _j .
_ij , and the bus bars 12 _i and 12 _j and the bus bar 15 of the load 14 are connected by power transmission lines 16 _i and 16 _j , respectively.

Step S11: The computer 1 periodically activates the program using a timer, or the program is activated by an on/off operation such as a disconnector, and calculates the system status each time. In this respect, humans
This is different from a power flow calculation system in which the program is started from an input device such as a CRT display.

Step S12: Compare the system configuration one cycle ago with the current system configuration to determine whether there has been a change.

Step S13: If there is no change in the previous step S12 as in ^FIG .
Y ^-1 and voltage V of each bus 12 _i and 12 _j , phase angle θ
is input from the disk device 2. Therefore, the effort of calculating the Jacobian inverse matrix J ^-1 each time the program is started can be saved.

Step S14: Each bus line 1 that was in the previous step S13
The voltage V and phase angle θ of 2 _i and 12 _j are set as initial values for convergence calculation. Therefore, this can save the effort of calculating the initial value every time the program is started.

Step S15: As in the prior art, the active power correction amount ΔP and the reactive power correction amount ΔQ are set in the previous step S13 and calculated using the Jacobian inverse matrix J ^-1 .

Step S16: According to equation (1), the voltage correction amount ΔV and the phase angle correction amount Δθ are calculated from the inverse matrix J ^-1 and the correction amounts ΔP and ΔQ.

Step S17: This calculation result is used again as a correction amount, and steps S15 to S16 are repeated until Δθ and ΔV converge in the above equation (1).

Step S18: θ and ΔV in step S17
If the judgment result that the has converged is obtained, the convergence value θ,
Power flows P _ij and Q _ij of the branches of the power system diagram shown in FIG. 2B are calculated using V.

Step S21: In step S12, Fig. 2A
From the state of , transmission line 13 _ij between bus bars 12 _i and 12 _j
The switchgear _17i and _17j installed in the switchgear 17i and ^17j are opened and the state shown in FIG. The matrix, ie, the Y ^-1 matrix, is partially modified in accordance with the configuration change.

Step S22: Initial values of the voltage V ₁ and phase angle θ of each bus 12 _i and 12 _j are calculated based on the system configuration after the configuration change.

Hereafter, steps S15 to S17 are repeated, and finally, in step S18, the active power P _ij of the branch power flow is
Calculate Q _ij .

Next, we will give a detailed explanation of the power flow calculation program shown in Figure 4. In this program, the calculation of the J ^-1 matrix is not included in the convergence calculation loop, and the J ^-1 matrix is not dependent on V and θ. Can calculate. That is, by dividing the systematic equation into P-θ and Q-V, it is replaced by the following equation.

P=Yθ+ΔP Q=YV+ΔQ (3) where Y is the admittance matrix, ΔP is the active power correction amount, and ΔQ is the reactive power correction amount.

If we apply the Newton-Raphson method to equation (3), we get
At the n-th step, θ _o+1 = Y ^-1 (P _set - ΔP _o ) Vo ₊₁ = Y ^-1 (Q _set - ΔQ _o ) (4). In this way, the J ^-1 matrix becomes P-θ
It is sufficient to use Y ^-1 for both the loop and the Q-V loop,
It is uniquely determined from the system configuration and can be calculated before the convergence calculation.

Also, if the system configuration does not change, the
You can use the Y ^-1 matrix as is.

Next, when the system configuration changes, J ^-1 matrix, that is, Y ^-1
All you have to do is partially modify the matrix. In other words, since the system configuration has changed, it is necessary to recalculate the Y ^-1 matrix, but since calculating the inverse matrix takes time, by using Kron's formula, we can calculate the inverse matrix from one cycle before. Partial correction.

For example, in Fig. 4, when Brante i-j is released, if the inverse matrix of admittance matrix Y before release is Y ^-1 (=zkl'), then zkl'=zkl-Z _ik -Z _jk ) (Z _il −Z _jl /−1/yij+(Z _ij
+Z _jj +2Z _ij )…(5) Here, k, l=1, 2,...N. In this way, the Jacobian inverse matrix J ^-1 is modified using equation (5).

Such correction calculations are performed in step S21, and then initial value calculations are performed in step S22, and the following processing is performed in the same way as when there is no change in the system configuration.

In the above embodiment, the load flow calculation program is set as one task and is started periodically, but the above program is executed as two tasks.
It is also possible to divide the process into two tasks.
In other words, Y ^-1 for one task due to a change in system configuration.
One task is to correct the matrix and calculate the initial values of V and θ, and another task is to periodically input the Y ^-1 matrix and the initial values of V and θ from the disk device and perform convergence calculations. The first task is activated when the system configuration changes, and the second task is activated periodically by a timer. In this way, the calculation time can be further reduced.

As described above, according to the present invention, there is no need to calculate the J ^-1 matrix for each repetition number, and if the system configuration does not change, the V ^-1 matrix of one cycle before can be referred to as is. The calculation time can be significantly reduced.

Also, when the system configuration changes, instead of calculating a new J ^-1 matrix,
All you need to do is partially modify the J ^-1 matrix.

Furthermore, if the system configuration does not change, the values of V and θ one cycle before are referenced and used as the initial values for convergence calculation, so there is no need to calculate the initial values.

By the way, by partially correcting the J ^-1 matrix, J ^-1
Matrix calculation takes about 1/N (N is the number of nodes) of the calculation time of the conventional case. For example, the execution time for a conventional program is 24 [sec] (that is, the preprocessing time T _pre in step S2 in Fig. 3 is 2 [sec],
The initial value calculation time T _iot of step S3 is 1 [sec], the correction calculation time T _hs of step S4 is 0.5 [sec], the J ^-1 matrix calculation time T _jac of step B5 is 3 [sec],
The voltage of S6, the phase angle calculation time T _vt is 0.5 [sec], the P-Q flow calculation time T _PQ of step S7 is 1 [sec],
Assume that the number of repetitions N _R is 5 and the number of nodes N = 50)
On the other hand, the program execution time in the case of the present invention is approximately 8.1 times when there is a change in the system configuration.
[sec] and there is no change in the system configuration 4.2
[sec] (i.e. in step S13 of Fig. 4)
J ^-1 matrix setting time T _jset is 0.1 [sec], step S14
Since the initial value of the set time T _ioset is 0.1 [sec] and the number of repetitions is N _R =1).

As described above, according to the present invention, in order to realize real-time processing of a training simulator compared to the conventional method, calculation of J ^-1 is not necessary when there is no change in the system configuration, and when there is a change. However, by making partial corrections instead of calculating J ^-1 all at once, it is possible to display periodic power flow calculation results faster than before.

[Brief explanation of the drawing]

Fig. 1 is a system diagram showing the hardware of a training simulator to which the present invention can be applied, Fig. 2 is a system diagram showing changes in system configuration, and Fig. 3 is a schematic flowchart of a conventional power flow calculation program. FIG. 4 is a flow chart showing a schematic flowchart of a tidal flow calculation program showing an embodiment of the training simulator according to the present invention. 1...Digital computer, 2...Magnetic disk device, 3...System typewriter, 4...Operation console, 5
... Card reader, 6... Line printer, 7... System monitoring board, 11i, 11j... Generator, 12i, 12
j, 15...Bus bar, 13ij, 16i, 16j...Power transmission line, 14...Load, 17i, 17j...Switching device.

Claims (1)

[Claims] 1. The configuration of a simulated power system is changed by inputting data from the console, and a computer is used to perform convergence calculations to obtain the convergence values of the bus voltage and phase angle in the system when the configuration changes. For training purposes, the power flow is calculated using the converged voltage value and phase angle, and the power flow when the system configuration changes is displayed and simulated on the system monitoring panel. When performing power flow calculations in the simulator, it is determined whether there is a change in the system configuration, and if there is no change, the Jacobian inverse matrix calculated and stored at the previous power flow calculation stage is used as is, or if there is a change, the Jacobian inverse matrix calculated and stored before the change is used. The initial value is obtained by partially modifying the Jacobian inverse matrix using the admittance of the part where the system configuration has changed, and based on this initial value, the calculation of the correction amount and the phase angle and voltage calculation are performed based on the phase angle. A training simulator characterized in that the calculation is repeated by a computer until the voltage converges, and the convergence value is used to calculate and display the active power and reactive power of the power flow of the branch.